Date: Tue, 23 Nov 2010 23:21:39 +0800 From: Adrian Chadd <adrian@freebsd.org> To: Larry Vaden <vaden@texoma.net> Cc: freebsd-mips@freebsd.org Subject: Re: does anyone have a recommendation for a solid FreeBSD-based 802.11b/g/n AP package for MIPS-based APs and CPEs? Message-ID: <AANLkTi=-tiwQ24KK8DV68RmYxbAk_0qHcpp=wxiaUg1R@mail.gmail.com> In-Reply-To: <AANLkTimibh2z2h%2BBJitxmPtguv2TZyeQYNMajs_SQtz7@mail.gmail.com> References: <AANLkTik_vvgQfhh_K-OecR-CE8nBWu=A3Kuvt8BYdXSZ@mail.gmail.com> <AANLkTikTRMMZHFjFOo-tWGgJ1=EVu0e2pMR8ScVVNfs9@mail.gmail.com> <AANLkTi=WhMTFK=eLaZb2DNG0NE0HmUBiQqfw1_6E29q1@mail.gmail.com> <AANLkTi==En8kS6H_7sxTGFMrEO3KLAbQwgMWsAwb9A2w@mail.gmail.com> <AANLkTimibh2z2h%2BBJitxmPtguv2TZyeQYNMajs_SQtz7@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On 23 November 2010 23:01, Larry Vaden <vaden@texoma.net> wrote: > You have been a very productive fellow while taking exams! Hah! :) I explicitly didn't do any of this during exams! > What can you tell us about how the stuck beacon problem is handled > (and indeed if the problem remains in the development code branch)? "stuck beacon" is a symptom of many, many potentially inter-related causes. I've fixed some of the code related problems that upset the AR9160's I was using. I can't fix RF issues, nor can I fix issues that require detailed documentation about the behaviour of the radio. I have almost no documentation about the atheros stuff and certainly nothing at all about the radio. All I really have is ath5k/ath9k and some helpful contacts that answer the (frequent!) questions I have. I haven't yet sat down and tried to use the AR9280 I have here in any real way; I'm specifically ignoring it until I've finished off the 11n TX path. I have to go over the ath9k code to see what bits we're missing that could impact stability - there's definitely calibration and TX power related code that we don't have in our HAL. Rui had problems with the AR9280 stability and I'm thus focusing on the chipset that -is- closer to working in a useful way (AR9160), even if it's "old" and "unsupported" today. > Will Rogers said "Everybody is ignorant, only on different subjects." > and I'm ignorant about TX. Well, in summary - the TX descriptor format has changed between chipsets (ie, what fields mean in the TX descriptors.) The 11n chipsets have different expectations for programming TX rates, rts/cts stuff, etc. There's also extra stuff to care about for handling 11n TX A-MPDU which I'm just going to ignore for the first pass. I've been slowly figuring out what's going on by reading ath9k/ath5k source and asking a few atheros people some questions but it's taking (a lot of) time. That said, the relevant ath5k/ath9k people have been very, very helpful. :-) I'm short on free time at the moment and I do need a bit of a break from concentration, so there's no real timeline for: * merging my platform work back into -head; * finishing the AR724x so (in theory) the Bullet M2/M5 can boot FreeBSD; * finishing up the if_ath TX path code so it works for both legacy and 11n chipsets "right enough" to trust it; * even thinking about merging any of this initial 11n stuff from Rui/Sam into -HEAD. I'll post an update on -mips when I've hacked together the AR724x support to test. Adrian
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?AANLkTi=-tiwQ24KK8DV68RmYxbAk_0qHcpp=wxiaUg1R>